Fix metadata schemas on assert_equals

Author: benjeffery

python/CHANGELOG.rst

@@ -19,13 +19,16 @@
 - Add ``TreeSequence.mutations_edge`` which returns the edge ID for each mutation's
   edge. (:user:`benjeffery`, :pr:`3226`, :issue:`3189`)
 
-
 **Bugfixes**
 
 - Fix bug in ``TreeSequence.pair_coalescence_counts`` when ``span_normalise=True``
   and a window breakpoint falls within an internal missing interval.
   (:user:`nspope`, :pr:`3176`, :issue:`3175`)
 
+- Fix metadata schemas that are equal but have different byte representations not being equal
+  when using ``TableCollection.assert_equals`` and ``Table.assert_equals``.
+  (:user:`benjeffery`, :pr:`3246`, :issue:`3244`)
+
 **Breaking changes** 
 
 - ``ltrim``, ``rtrim``, ``trim`` and ``shift`` raise an error if used on a tree sequence

python/tests/test_tables.py

@@ -1401,6 +1401,21 @@ def test_metadata_schema(self, table_5row):
                 table_5row.assert_equals(table2)
             table_5row.assert_equals(table2, ignore_metadata=True)
 
+    def test_degenerate_metadata_schema(self, table_5row):
+        if hasattr(table_5row, "metadata_schema"):
+            table_5row.clear()
+            table1 = table_5row.copy()
+            table2 = table_5row.copy()
+            table1.ll_table.metadata_schema = (
+                '{"codec": "json", "properties": '
+                '{"A": {"type": "integer"}, "B": {"type": "number"}}}'
+            )
+            table2.ll_table.metadata_schema = (
+                '{"codec": "json", "properties": '
+                '{"B": {"type": "number"}, "A": {"type": "integer"}}}'
+            )
+            table1.assert_equals(table2)
+
     def test_row_changes(self, table_5row, test_rows):
         for column_name in test_rows[0].keys():
             table2 = self.table_class()
@@ -4217,6 +4232,17 @@ def test_ignore_reference_sequence(self, t1, t2):
             t1.assert_equals(t2)
         t1.assert_equals(t2, ignore_reference_sequence=True)
 
+    def test_degenerate_metadata_schema(self, t1, t2):
+        t1._ll_object.metadata_schema = (
+            '{"codec": "json", "properties": '
+            '{"A": {"type": "integer"}, "B": {"type": "number"}}}'
+        )
+        t2._ll_object.metadata_schema = (
+            '{"codec": "json", "properties": '
+            '{"B": {"type": "number"}, "A": {"type": "integer"}}}'
+        )
+        t1.assert_equals(t2)
+
 
 class TestTableCollectionMethodSignatures:
     tc = msprime.simulate(10, random_seed=1234).dump_tables()

python/tskit/tables.py

@@ -462,9 +462,18 @@ def assert_equals(self, other, *, ignore_metadata=False):
                 f"other={other.num_rows}"
             )
 
+        # We can reach this point if the metadata schemas byte representations
+        # differ when the decoded schema is the same
+        if (
+            not ignore_metadata
+            and self.ll_table.metadata_schema != other.ll_table.metadata_schema
+            and self.metadata_schema == other.metadata_schema
+        ):
+            return
+
         raise AssertionError(
             "Tables differ in an undetected way - "
-            "this is a bug, please report an issue on gitub"
+            "this is a bug, please report an issue on github"
         )  # pragma: no cover
 
     def __eq__(self, other):
@@ -2411,7 +2420,7 @@ def assert_equals(self, other, *, ignore_timestamps=False):
 
         raise AssertionError(
             "Tables differ in an undetected way - "
-            "this is a bug, please report an issue on gitub"
+            "this is a bug, please report an issue on github"
         )  # pragma: no cover
 
     def add_row(self, record, timestamp=None):
@@ -3277,9 +3286,18 @@ def assert_equals(
                 other.provenances, ignore_timestamps=ignore_timestamps
             )
 
+        # We can reach this point if the metadata schemas byte representations
+        # differ when the decoded schema is the same
+        if (
+            not ignore_metadata
+            and self._ll_object.metadata_schema != other._ll_object.metadata_schema
+            and self.metadata_schema == other.metadata_schema
+        ):
+            return
+
         raise AssertionError(
             "TableCollections differ in an undetected way - "
-            "this is a bug, please report an issue on gitub"
+            "this is a bug, please report an issue on github"
         )  # pragma: no cover
 
     def __eq__(self, other):